Delete the constant time string comparison. It's slow and shows up in
profiles, nobody else does it, and the spec text no longer recommends
doing it. See bug for discussion and details.

Also clean up some naked returns while I'm here (noted during review).

Fixes golang/go#19238

Change-Id: I344c5766c5d97bbcf01eab0624097941591ce00f
Reviewed-on: https://go-review.googlesource.com/37394
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Tom Bergan <tombergan@google.com>

The x/net package is currently broken for Go 1.9 (#19051) so
I am unable to use trybots for x/net/http2.

This disables the tests for the broken stuff and makes things compile
at least, so x/net trybots aren't broken for others.

Updates golang/go#19051

Change-Id: I67401d7ad32d855e99a417545328eb4e803287cc
Reviewed-on: https://go-review.googlesource.com/37401
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Martin Möhrmann <moehrmann@google.com>

This change drops a workaround for golang.org/issues/13372.

The compiler in Go 1.6 or above handles the "conversion of a
unsafe.Pointer to uintptr when calling syscall.Syscall" case
correctly.

Also fixes a typo.

Change-Id: I0433f3b8f75b34437aad91c6e8cf103e884d2a83
Reviewed-on: https://go-review.googlesource.com/37171Reviewed-by: Ian Lance Taylor <iant@golang.org>

This change drops a workaround for golang.org/issues/13372.

The compiler in Go 1.6 or above handles the "conversion of a
unsafe.Pointer to uintptr when calling syscall.Syscall" case
correctly.

Change-Id: I5b45094a78f15a631da277bbea0cb79fde25bb2a
Reviewed-on: https://go-review.googlesource.com/37170Reviewed-by: Ian Lance Taylor <iant@golang.org>

This package doesn't work with Go 1.9 or above.
The bypass surgery for grafting the net and internal/poll packages of
Go 1.9 starts with disabling this packge.

Updates golang/go#19051.

Change-Id: I82d742d267f155cf440884e845b24402a5a5bfdf
Reviewed-on: https://go-review.googlesource.com/37034Reviewed-by: Ian Lance Taylor <iant@golang.org>

Update the list to revision 45a2bf8ef3e22000fbe4bfa5f9252db41d777001
(2017-01-18T01:04:06Z).

Change-Id: I39c865c3ac59f381a8926ca51b652d1e6a3a035d
Reviewed-on: https://go-review.googlesource.com/35850Reviewed-by: Nigel Tao <nigeltao@golang.org>

Split off a new SOCKS5 connect() method from Dial.

connect() takes an existing connection to a socks5 server, and
commands the server to extend that connection to a given target
address and port.

Change-Id: I5dbba58a67a0d884bda3d3ac194dc18bdebe74ab
Reviewed-on: https://go-review.googlesource.com/36643Reviewed-by: Adam Langley <agl@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: Ic8192f74337e021a5d72feb07b9648619d778235
Reviewed-on: https://go-review.googlesource.com/36673Reviewed-by: Matt Layher <mdlayher@gmail.com>
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Change-Id: Ia8bc87d0398951d7be74779279dca177f499d293
Reviewed-on: https://go-review.googlesource.com/36377
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Change-Id: I954d9a3f97a68c1366e917a4a430dd832227e132
Reviewed-on: https://go-review.googlesource.com/36376Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

This change adds the Marshal method of RouteMessage to make it
possible to exchange route messages between userspace processes and
the kernel for the manipulation of routing information base inside the
kernel.

Change-Id: I0cf2c1a391820f41eb9c5eac1c172598cb2e1533
Reviewed-on: https://go-review.googlesource.com/36077Reviewed-by: Ian Lance Taylor <iant@golang.org>

Fixes golang/go#18809

Change-Id: Ib1014f3ebe5a57dde30b4eaf287a2cbff3c1179c
Reviewed-on: https://go-review.googlesource.com/36118
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Tom Bergan <tombergan@google.com>

The tree's pretty inconsistent about single space vs double space
after a period in documentation. Make it consistently a single space,
per earlier decisions, and changes in go repository. This means
contributors won't be confused by misleading precedence.

This CL was generated with:

	perl -i -npe 's,^(\s*// .+[a-z]\.)  +([A-Z]),$1 $2,' $(git grep -l -E '^\s*//(.+\.)  +([A-Z])')

on top of copyright headers change in https://golang.org/cl/32878.

Follows https://golang.org/cl/20022.

Change-Id: I821e4a300122b4668aa31e12eaa914db615e5369
Reviewed-on: https://go-review.googlesource.com/32879Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

This reverts commit 67957fd0.

Updates golang/go#18567

Change-Id: I4a9da509eb95949d2e3ab08763274abf6706f6f8
Reviewed-on: https://go-review.googlesource.com/35270
Run-TryBot: Nigel Tao <nigeltao@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Benoit Sigoure <tsunanet@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Fixes golang/go#18538

Change-Id: Ic0627352f96ad5fa138633d1e1ccfaf76294d621
Reviewed-on: https://go-review.googlesource.com/35171
Run-TryBot: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: David Anderson <dave@natulte.net>
Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Parsing the templates can take a significant amount of
time(30-35ms) so it should not be done on the init phase.

Change-Id: I8f258b8028510e698a97b55faeac0d28a61b7b22
Reviewed-on: https://go-review.googlesource.com/21654Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Change-Id: Id87fcbe2d827f5b3ee3b7387e1bbf64287a397bf
Reviewed-on: https://go-review.googlesource.com/33811
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Change-Id: I4473e69bbb445ab27554f2dc625d0cbd79d7cfe5
Reviewed-on: https://go-review.googlesource.com/35072
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>

Also updates the package description.

Change-Id: I782a9d34d5a35aff92873efc6d511f95686cd5d6
Reviewed-on: https://go-review.googlesource.com/35071
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>

Oops: forgot to git commit these nits locally before uploading
https://go-review.googlesource.com/34984

Change-Id: I98995ea9f6ecdde731cd298ff54afe5047ae136e
Reviewed-on: https://go-review.googlesource.com/34985
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

I believe there were two bugs, both fixed by this CL.

* Previously, we checked stateHalfClosedRemote before waiting for the
  PUSH_PROMISE. However, the pushed stream is not created until the promise
  is written, so the stream may not have started yet, which means we'd see
  stateIdle instead of stateHalfClosedRemote.

* The push reponse handler cannot write the response until after we
  check the pushed stream state. Otherwise, the response might finish
  just before we check the stream state and we'll see stateClosed instead
  of stateHalfClosedRemote.

Test passes with -count 1000.

Fixes golang/go#18559

Change-Id: I61f62635957e061fba905a41dcb15cd4e563031a
Reviewed-on: https://go-review.googlesource.com/34984
TryBot-Result: Gobot Gobot <gobot@golang.org>
Run-TryBot: Tom Bergan <tombergan@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I61475922625ce8dbeee4180807c3f021a502edf7
Reviewed-on: https://go-review.googlesource.com/34731
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I468795cc9ab4d2350d20427e62676df5d89b2c3b
Reviewed-on: https://go-review.googlesource.com/34733
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

The existing implementation passes hostname:port to the ServerName
field of the client's TLS config. This is passed to the server
incorrectly as the ServerName in the client hello. This change adds a
function to strip the port from the host when passing it to the TLS
config.

Change-Id: I03714ffc7f21d87c375f8f07392ef02bbe76da66
Reviewed-on: https://go-review.googlesource.com/34728
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: Ia77cabda5ae189dd91c3e528b9769ba93581f00d
Reviewed-on: https://go-review.googlesource.com/34732
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I95161246bbb1a29490c0058593ee36bcb706e1d9
Reviewed-on: https://go-review.googlesource.com/34820
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

The "fake" jump conditions as well as the LoadExtension instructions
are now disassembled correctly. Therefore the workaround to reassemble
the disassembly is no longer necessary.

This simplification was annonced already in golang/go#18470.

Result of `go test -cover .` stays the same with this simplification.

$ go test -cover golang.org/x/net/bpf
ok  	golang.org/x/net/bpf	0.495s	coverage: 92.3% of statements

Change-Id: I3f9eb46148287c76059437b773b80c4c99eb5b53
Reviewed-on: https://go-review.googlesource.com/34951
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>

Change-Id: I54dbe1ec47e6020d1d34b4f55f1e94c7800f8c12
Reviewed-on: https://go-review.googlesource.com/34874
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Load extension in BPF is an overload of the load absolute
instruction with a 'negative offset'. Disassemble load
absolute instructions to load extension, if the offset
is 'negative', respectively > (0xFFFFFFFF - 0x1000).

Fixes golang/go#18469

Change-Id: I13af8f5af89ce26b13a35d2b06879dd6e1c7434e
Reviewed-on: https://go-review.googlesource.com/34771
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>

JumpNotEqual, JumpLessThan and JumpLessOrEqual are "fake" jump
conditions that don't appear in the machine code. Nevertheless
these instructions (jneq/jne, jlt, jle) are listed in the
specification and therefore they should be reconstructed from the
machine code.

Fixes golang/go#18470

Specification: https://www.kernel.org/doc/Documentation/networking/filter.txt

Change-Id: I9116b99056e379d89e71adc90516c6747d388e5d
Reviewed-on: https://go-review.googlesource.com/34772
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>

Change-Id: I08148b1a7fa5f57bcf71f4e49a698360d33a6f8e
Reviewed-on: https://go-review.googlesource.com/34872
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I00b3297b88142ddb3c0cd08391171ebb6e731550
Reviewed-on: https://go-review.googlesource.com/34751
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I842503c64b7b61913d0748cd5b5618543bcf198b
Reviewed-on: https://go-review.googlesource.com/34752
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

The API extends the old API but is intended to be backwards
compatible. The old test file serves as a test for this purpose.

I will send a CL to update the vendored mirror in core after
1.8 is out.

Change-Id: I538df20d42c213d9e038bdb0dad76b6cc132d4fd
Reviewed-on: https://go-review.googlesource.com/34770
Run-TryBot: Marcel van Lohuizen <mpvl@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>

strings.Join has two allocations:
(1) []byte intermediate
(2) final string conversion
The comma-separated keys are ultimately stored in hpack.HeaderField.Value
as a string so (2) is not wasteful. Because strings.Join is used so heavily
I assume its implementation is optimal - (1) is necessary.
Therefore, short of refactoring hpack.HeaderField's Value type, the status quo
seems optimal.

This TODO appears to have been copy-pasted from net/http/transfer.go
However, the two cases are different. transfer.go has two allocations
in addition to strings.Join: concatenation with literals and converting
string to bytes in io.WriteString.

Change-Id: I292203a6535dd2774f5bf45e7280b89aabb5d78d
Reviewed-on: https://go-review.googlesource.com/34626Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Current handling of WriteTimeout for http2 does not
extend the timeout on new streams. Disable the WriteTimeout
in http2 for 1.8 release.

Fixes test added in https://golang.org/cl/34723

Updates golang/go#18437

Change-Id: I366899fb4ff2e740610cad71e004141d092699a2
Reviewed-on: https://go-review.googlesource.com/34724Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Change-Id: I22b3781a4514b7e8a3778021c465822022b8bdc2
Reviewed-on: https://go-review.googlesource.com/34499Reviewed-by: Tom Bergan <tombergan@google.com>

Previously, we panic'd on sending WINDOW_UPDATE on the half-closed-local
state. However, the RFC allows sending WINDOW_UPDATE in this state, so
we should no panic.

Change-Id: I702b2d5ad525837d7b8c650b7a2ed3f16371be63
Reviewed-on: https://go-review.googlesource.com/34498
Run-TryBot: Tom Bergan <tombergan@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Updates golang/go#18326

Change-Id: Iaed30073bab10c7fc25132b388d7b2fc101293d5
Reviewed-on: https://go-review.googlesource.com/34494Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Writing the PUSH_PROMISE can happen concurrently with the request
handler, so we should pass the request handler a cloned Header..

Updates golang/go#18326

Change-Id: I918e7ffbc02459b9ede3226f4bd3c56a6a3b350f
Reviewed-on: https://go-review.googlesource.com/34493
TryBot-Result: Gobot Gobot <gobot@golang.org>
Run-TryBot: Tom Bergan <tombergan@google.com>
Reviewed-by: Chris Broadfoot <cbro@golang.org>